New Year’s Robolutions

The Year in Robotics: Reflecting on 2021 and Looking Ahead to 2022

The period between the Christmas holiday and the New Year often feels unique – a transitional phase. Known as “Romjul,” or “Dead Week,” in Norway, it’s a time for both quiet contemplation and, for some, navigating CES embargoes. Here at Actuator HQ, we’re balancing both aspects.

Over recent weeks, we’ve analyzed key trends within the robotics sector: delivery systems, warehouse and fulfillment processes, and food preparation technologies. We’ve also engaged with prominent figures in the field, including Matthew Johnson-Roberson of CMU Robotics, Rob Playter, CEO of Boston Dynamics, Brian Gerkey of OpenRobotics, and iRobot CEO Colin Angle.

Insights from MIT CSAIL Director Daniela Rus

This week, we invited Daniela Rus, Director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), to share her perspective. Her comprehensive responses for this final Actuator edition of 2021 are particularly insightful.

What was the most significant robotics, AI, or automation trend of 2021? The pandemic and resulting labor shortages underscored the vital role robots can play in the workplace. We’ve observed increased robot adoption in manufacturing and logistics, where autonomous systems deliver demonstrable value. However, fully autonomous vehicles, like robotaxis, remain a future prospect. Significant research progress has been made in developing safer and more capable robots, including advancements in soft robotics and machine learning-driven robotic intelligence.

In the realm of AI, awareness of the limitations of current solutions has grown. While deep neural networks are augmenting work across various fields, their reliance on massive, manually labeled datasets presents a challenge. Data quality is paramount; biased or flawed data leads to equally flawed system performance. Furthermore, these systems often function as “black boxes,” hindering true understanding of their internal processes. Robustness remains a concern, as trained models can be unstable, and these systems primarily engage in pattern matching rather than deep reasoning. The research community is actively addressing these issues.

Looking Forward: Predictions for 2022

What does 2022 hold for these categories? As the world continues to evolve rapidly, robots and AI have the potential to unlock human potential, both individually and collectively. The past 60 years have focused on industrial robots executing complex tasks in controlled environments. The next 60 years will see robots integrated into human-centric environments, assisting with physical tasks.

Past industrial robots were largely modeled on the human form. Future robots will draw inspiration from the animal kingdom, exhibiting greater form and diversity, and adapting to our built environment. Materials will also shift, moving away from hard plastics and metals towards naturally available or engineered materials like wood, paper, ice, or even food.

We will also see renewed focus on AI privacy and sustainability. Progress relies on decades-old ideas enhanced by data and computation. New ideas and funding are crucial. Understanding the carbon footprint of machine learning and pursuing Sustainable AI will become increasingly important.

AI can optimize activities to mitigate climate change impacts – reducing technology’s energy consumption, improving transportation efficiency, monitoring deforestation, preserving biodiversity, and ensuring food security. However, AI systems require substantial energy. Training a large deep-learning model generates significant carbon dioxide emissions. Developing simpler models is essential to reduce AI’s environmental impact.

The expansion of robotics, automation, and AI has the potential to improve lives, but may also displace human workers. Efforts to anticipate and address potential economic inequality are vital.

Professor Rus’ insights are invaluable. This year has been remarkable for robotics, with predictions becoming reality faster than anticipated, largely driven by the pandemic. Addressing climate change and the potential displacement of human jobs remain critical challenges.

Roboticists often highlight the dull, dirty, and dangerous jobs they aim to replace. However, the broader societal impact on human workers requires careful consideration. While companies currently struggle to fill roles – a problem automation can address – much of the technology still requires human collaboration. As technology advances, will we inadvertently leave large segments of the population behind by deeming their work “unskilled”?

Key Categories in Robotics – A Year in Review

The following categories have been central to our coverage this past year:

Warehouse/fulfillment
Transportation
Food delivery/prep
Farming/agriculture
Home
Medical/surgical
Manufacturing
Construction

This is just the beginning. For example, Petra, a drilling company, recently raised $30 million to develop robotic rock-boring technology for infrastructure projects. The list of industries robotics is poised to disrupt continues to grow.

Industry News and Updates

As we approach the New Year, a few recent developments deserve attention. Meituan is expanding drone-based food delivery in Shenzhen, having completed 19,000 deliveries to 8,000 customers during a two-year pilot program. Relatively lenient regulations in Shenzhen have facilitated this progress.

China’s Ministry of Industry and Information Technology has outlined a five-year plan to boost robotic adoption, targeting a 20% annual revenue increase, with a focus on manufacturing.

Happy New Year! We hope you’ve navigated the challenges of the past year successfully.